Rescue.ETH
Autonomous system that prevents Aave liquidations using fast, cross-chain execution
Project Description
Rescue.ETH is an autonomous liquidation-prevention system designed to protect Aave positions before they are liquidated. Instead of relying on alerts or manual intervention, Rescue.ETH continuously monitors a user’s Health Factor off-chain and reacts instantly when risk thresholds are breached.
When a position becomes unsafe, an off-chain keeper triggers a deterministic execution flow that uses direct LI.FI contract calls to source liquidity cross-chain and repay or supply on Aave v3 on the destination chain. The system is optimized for speed and reliability, avoiding vault dependencies and minimizing execution latency. Configuration is managed via ENS-based records, allowing flexible updates without redeployments. The full rescue flow is tested using Tenderly forks to simulate real liquidation scenarios.
Rescue.ETH acts as an active defense layer for DeFi lending positions, turning liquidation prevention from a manual reaction into an automated, cross-chain safety mechanism.
How it's Made
Rescue.ETH is built as a speed-first, off-chain automated rescue system for DeFi lending positions. The core architecture uses a Node.js keeper that continuously monitors Aave v3 positions and computes the Health Factor in real time using on-chain reads. We intentionally keep this logic off-chain to minimize latency and avoid on-chain automation constraints.
When the Health Factor drops below a configurable threshold, the keeper triggers a deterministic execution module (execute.ts). This module performs direct LI.FI contract calls to source and bridge liquidity across chains and execute the rescue in a single flow, repaying or supplying assets on the destination Aave v3 / L2Pool. No vaults or intermediate custody are used, reducing complexity and execution risk.
Configuration such as thresholds, target chains, and contract addresses is managed via ENS text records, allowing behavior updates without redeploying contracts. The entire system is tested using Tenderly forks, enabling realistic liquidation simulations and safe iteration on cross-chain execution paths. This architecture prioritizes reliability, speed, and real-world deployability over demo-only abstractions.
